[Meanski]

Item 9 of the parity roadmap.

• Export a table to CSV, JSON, or SQL (INSERT statements in the connection's dialect). You choose where to save in a native dialog. Exports stop at 200,000 rows and say so if the table is bigger.
• Import CSV, chosen in a native dialog, into a table:
  • Header names match the table's columns case-insensitively. Unknown or repeated columns, or rows with the wrong number of fields, are refused before anything is written.
  • Empty fields import as NULL.
  • All rows go in one transaction, as batched multi-row INSERTs, so a failure part-way leaves the table untouched.
  • Limits: 50 MB and 200,000 rows.
• The CSV parser handles quoted fields, embedded commas, quotes and newlines, CRLF line endings and a BOM. The writers quote and escape per format.
• Main only touches files the user picked in a native dialog, and the renderer never supplies a path.
• The PostgreSQL and MySQL connectors moved out of database.ts into dbPostgres.ts and dbMysql.ts (shared types in dbTypes.ts), keeping the schema fixes from feat: edit, insert and delete database rows #18 and feat: ER diagram of tables and foreign keys #19.

Tests cover the CSV parser and writers, header mapping, batching, transaction rollback, limits, and the export/import handlers for both dialects. The features page is updated.

[cloudflare-workers-and-pages]

Deploying noxed with    Cloudflare Pages

Latest commit:	60686bb
Status:	✅  Deploy successful!
Preview URL:	https://0dab4172.noxed.pages.dev
Branch Preview URL:	https://feat-db-import-export.noxed.pages.dev

View logs

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

• Configuration used: Repository: Meanski/noxed/.coderabbit.yaml
• Review profile: CHILL
• Plan: Advanced
• Run ID: 3280d1bf-0e74-4bc6-bc3c-e1e859d1aa84

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

• Autopilot · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[Copilot]

🟡 Changes recommended

CSV edge cases, IPC validation, table-context loss, non-round-trippable SQL values, and unbounded export memory usage remain unresolved.

4 open findings

• Export size is unbounded in memory · New
• Prototype properties bypass format validation · New
• Quoted empty CSV fields are incorrectly dropped · New
• Table context is lost during query refresh · New

What changed in this PR

Adds native-dialog table export to CSV, JSON, or SQL and transactional CSV import for PostgreSQL, MySQL, and MariaDB.

Changes:

• Adds import/export UI and IPC APIs.
• Implements CSV parsing, format writers, limits, and transactional batching.
• Extracts database connectors into dialect-specific modules and adds tests/documentation.

File	Description
ResultsTabsBar.tsx	Adds table import/export actions.
ExportTableModal.tsx	Adds export format selection.
DatabaseExplorer.tsx	Integrates import/export workflows.
DatabaseExplorer.test.tsx	Tests renderer workflows.
harness.tsx	Mocks transfer APIs.
preload/​index.ts	Exposes transfer IPC calls.
preload/​index.d.ts	Types transfer APIs.
dbTypes.ts	Defines shared connector types and batching.
dbTransfer.ts	Implements CSV and export serialization.
dbPostgres.ts	Extracts PostgreSQL connector and imports.
dbMysql.ts	Extracts MySQL/MariaDB connector and imports.
database.ts	Adds native file transfer handlers.
dbTransfer.test.ts	Tests parsing and writers.
database.test.ts	Tests handlers and transactions.
docs/​features.html	Documents table transfers.

🧠 Review effort: Balanced

---

Give feedback about Copilot approvals in this survey to enter a drawing for a $150 gift card.

[Copilot]

🟡 Changes recommended

Nontransactional MySQL tables and several serialization edge cases can cause partial imports or corrupted exports.

4 open findings

• Failed exports can delete the existing destination file · New
• PostgreSQL arrays are serialized as invalid JSON strings · New
• Binary values are exported as base64 text · New
• Final record with two empty fields is dropped · New

4 resolved since last review

• Export size is unbounded in memory
• Table context is lost during query refresh
• Quoted empty CSV fields are incorrectly dropped
• Prototype properties bypass format validation

Previously missed (2)

In code that hasn't changed since last review

MySQL nontransactional tables break import atomicity

src/​main/​ipc/​dbMysql.ts:103

A MySQL transaction does not provide the promised all-or-nothing import for nontransactional tables such as MyISAM: earlier batches remain inserted when a later batch fails and rollback() is ineffective. Check the target table's storage engine and reject imports unless it supports transactions, or use an import strategy that is atomic for that engine.

Case-insensitive column mapping collapses distinct PostgreSQL columns

src/​main/​ipc/​dbTransfer.ts:159

This map silently collapses PostgreSQL columns that differ only by case, which are legal when quoted (for example, Foo and foo). A CSV header can then be routed to whichever column appeared last and import data into the wrong column. Detect case-folding collisions and either prefer an exact-case match or reject an ambiguous header before inserting.

🧠 Review effort: Balanced

---

Give feedback about Copilot approvals in this survey to enter a drawing for a $150 gift card.

[Copilot]

🟡 Changes recommended

SQL exports can corrupt supported values, and imports are not atomic for non-transactional MySQL/MariaDB tables.

2 open findings

• Preserve PostgreSQL JSON scalar values during export · New
• Reject ambiguous case-insensitive PostgreSQL column names · New

4 resolved since last review

• Binary values are exported as base64 text
• PostgreSQL arrays are serialized as invalid JSON strings
• Failed exports can delete the existing destination file
• Final record with two empty fields is dropped

Previously missed (2)

In code that hasn't changed since last review

Prevent non-transactional MySQL imports from claiming atomicity

src/​main/​ipc/​dbMysql.ts:103

A transaction does not guarantee rollback for non-transactional MySQL/MariaDB table engines such as MyISAM (and some Aria configurations). If a later row or batch fails, earlier inserts remain even though rollback() is called, violating the advertised all-or-nothing import behavior. Check the target table engine and refuse imports when transactions are unsupported, or use an import strategy that guarantees atomicity.

Preserve PostgreSQL special floating-point values during export

src/​main/​ipc/​dbTransfer.ts:147

This silently exports PostgreSQL's valid NaN, Infinity, and -Infinity floating-point values as SQL NULL, changing the data when the script is replayed. Emit PostgreSQL's accepted quoted special literals instead; MySQL need not receive these values because it does not support storing them as ordinary floating-point values.

🧠 Review effort: Balanced

---

Give feedback about Copilot approvals in this survey to enter a drawing for a $150 gift card.

[Copilot]

🔵 Needs a closer look

Nontransactional MySQL tables violate import atomicity, and SQL serialization can corrupt or reject supported values.

0 open findings

2 resolved since last review

• Reject ambiguous case-insensitive PostgreSQL column names
• Preserve PostgreSQL JSON scalar values during export

Previously missed (4)

In code that hasn't changed since last review

Add tests covering multi-batch imports and rollback

src/​main/​ipc/​__tests__/​database.test.ts:264

This test imports only two rows, and the MySQL test below does the same, so neither crosses the 60,000-bind-parameter batch boundary despite the PR description claiming batching coverage. A regression in second-batch placeholder numbering or rollback after an earlier batch would pass. Make the batch limit injectable for tests or construct an import that forces at least two batches and fails the latter.

Reject nontransactional engines to preserve atomic imports

src/​main/​ipc/​dbMysql.ts:100

beginTransaction() does not make inserts atomic for nontransactional MySQL/MariaDB engines such as MyISAM or MEMORY; rollback is ignored and earlier batches remain written. This contradicts the all-or-nothing import contract for supported connections. Check the target table engine's transaction capability before inserting and reject unsupported tables (or explicitly narrow the feature contract).

Generate SQL literals compatible with NO_BACKSLASH_ESCAPES

src/​main/​ipc/​dbTransfer.ts:146

This escaping depends on MySQL's default backslash mode. With the valid NO_BACKSLASH_ESCAPES SQL mode, an original value such as C:\temp is replayed with two backslashes, so the SQL export silently changes data. Emit mode-independent literals (for example, encoded hex text) or include an explicit compatible SQL mode in the script.

Preserve PostgreSQL special values and typed object serialization

src/​main/​ipc/​dbTransfer.ts:177

This generic normalization loses valid PostgreSQL values. pg returns float NaN/Infinity as non-finite numbers, which become NULL, and object-valued types such as interval and point are JSON-stringified by plain() into text those columns cannot parse. The resulting SQL either corrupts data or cannot be replayed; preserve column type metadata and serialize/cast these values to PostgreSQL-compatible text.

🧠 Review effort: Balanced

---

Give feedback about Copilot approvals in this survey to enter a drawing for a $150 gift card.